Digital activity trackers have the potential to either revolutionize wellness data collection, or to end up collecting dust in a desk drawer. However, researchers at Partners HealthCare in Boston are undertaking a project, the building of an "engagement engine," that utilizes machine learning to help quantify what will encourage tracker users to keep on using them in pursuit of better health.

"We don't want these trackers to sit in the drawer, and that was where this effort was born," principal investigator Kamal Jethwani, M.D., senior director of connected health innovation at Partners, told Health Data Management. "If we can get you to consistently wear a tracker, you will walk 18 percent more. There are studies that have demonstrated that. We know that already. The problem is people don't use it long enough to see that effect."

Also See: Vendors, Organizations Start Cranking Out Apple Watch Apps

Jethwani believes that carefully examining the question of which tracker of the multitudes available might work best for certain types of users, and which activities are most appealing to them, could provide the spark to make a more healthy lifestyle an intrinsically rewarding endeavor.

He and his colleagues have received a $468,000 grant from the Robert Wood Johnson Foundation to create the engagement engine. The researchers will choose an initial cohort of 20 patients from three community clinics in the Boston area to create the engine. The study's target audience, he said, is broken into three segments: people who aren't doing anything to stay active, people who have done something in the past but don't do it anymore, and people who think they are doing something but really aren't (that is, those who feel tired at the end of the day due to general busy-ness, but whose activity level is really rather low, such as the 2,000 steps one may walk in a normal day).

After the initial phase of the study, in which the researchers will try to match the correct tracker to the cohort's individuals, Jethwani estimates the project's algorithm creation phase will start in June and take four to five months to finish.

"It's really an engine that will take into account your motivational style, your clinical background and history, and your technology ownership and readiness," he said. "We will use those three things to recommend an activity program for you, basically  – we would choose from 20 different tracking mechanisms to give you a tracker suitable for you, and then choose from 10 to 20 different ways in which you could track your activity and recommend one that is suitable for you, then watch you with that. Then, based on whether you are successful or not, and what outcomes you had, that data goes back to the machine learning algorithm also. That improves the algorithm over time."

Deborah Bae, senior program officer at the RWJF, said the organization funded the study in an effort to better understand how best to impart an improved sense of well-being to people outside clinical settings. Both Bae and Jethwani say a bottom-up approach to engagement will be crucial.

"I think there's a very strong belief in our C-suite that in five to 10 years, Partners HealthCare will no longer be in the illness management space – we will be in the wellness management space," Jethwani said, cautioning that executive-level interest should not translate into corporate-dictated programs.

"Patient engagement strategies cannot be created without engaging the patient in the first place," he said. "And I see tons and tons of people creating these engagement strategies where the C levels get together and come up with it, and I struggle a little bit with that. A patient engagement strategy is one where we are really in touch with patients, asking them what they need, figuring out from their perspective what is important to them and asking them how they want to be engaged in their care and with the hospital system. A lot of our current measures, even the HCAHPs we measure ourselves on, do a terrible job of measuring how 'engaged' the patient is."

Bae said the Partners study and a previously-funded project, the Health Data Exploration project coordinated by two University of California campuses, are just two early examples of not only discovering what kind of patient data may prove clinically relevant and useful to the user, but also demarcate a new era wherein clinicians will have to think of those they care for as consumers first and patients second.

Being able to quantify what works in keeping people healthier will be a vital step in the gargantuan task of changing the entire health system from the episodic model to an ongoing monitoring of health status, according to Bae, who would like to envision a day when a physician can prescribe a proven exercise regimen and be reimbursed for it instead of a medication.

"Hopefully, from that project and this project, we'll start to get really smart people in the room to start moving the needle to change prescription behavior and reimbursement models," she said. "There's a whole system we have to change behind it."

Register or login for access to this item and much more

All Health Data Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access